Intermarriage and Mother-Child Relationships

Abstract

Research indicates that when adult children marry, ties to parents weaken. Yet less is known about how spousal characteristics, and specifically, spouse’s race or ethnicity, affect ties to the family of origin. This paper uses data from the National Longitudinal Study of Adolescent to Adult Health to ask how interracial/ethnic marriage, compared to same-race/ethnicity marriage, is associated with ties to mothers among young adults in the United States. Results indicate that offspring who are intermarried differ little in their relationships to mothers compared to those who married same-race/ethnicity partners. However, findings from this study suggest that intermarriage may have greater consequences for some groups, such as Blacks, compared to other racial/ethnic groups. Overall, the results highlight how intermarriage has a relatively limited effect on offspring relationships with mothers and suggest a role for future research that examines how ties to parents during adolescence may shape partner choices during adulthood.

The marriage of adult children presents a potential shift in parent-child relationships. With marriage, commitments to family members are renegotiated and re-appraised in light of new members who must be brought into the family fold. Historians argue that marriage once strengthened parent-child bonds (Coontz, 2005). However, recent research finds that married offspring spend less time with parents than those who are single (Musick & Bumpass, 2012; Sarkisian & Gerstel, 2008). What is less known, however, is whether certain spousal traits also influence offspring relationships to parents. In particular, scholars often overlook how marriage to someone of a different race/ethnicity than one’s own shapes ties to the family of origin. Yet interracial/interethnic marriages – intermarriages – have more than doubled over the past 30 years, from 6.7% of all new marriages in 1980 to 15.1% in 2010 (Qian & Lichter 2007; Qian & Lichter, 2011; Wang, 2012). A parallel trend suggests that public opposition to intermarriage has also declined steadily since the 1950s (Newport, 2013). Young adults today are much more likely to have liberal attitudes about intermarriage than those in the past (Pew Research Center, 2010; Wang, 2012).

Yet, lingering cohort differences in attitudes about intermarriage could negatively impact intergenerational ties when offspring marry outside of their racial/ethnic group. In 2009, 81% of non-Hispanic Whites born between 1984 and 1995 (“Millennials”) said it would be fine if a family member married someone of a different race/ethnicity, whereas only 52% of non-Hispanic Whites born between 1949 and 1963 (“Baby Boomers”) agreed (Wang, 2012: 38). Thus, despite the growing public acceptance of intermarriages, cohort differences suggest that parents may still object to their children marrying outside of the family’s race/ethnicity.

Although interviews with intermarried couples suggest that they may have weaker ties to parents than those in same-race/ethnicity marriages (Nemoto, 2009; Killian, 2013; Root, 2001; Rosenblatt et al., 1995), no large-scale nationally representative studies exist to support these findings. To address this gap, this paper assesses how intermarriage compares with same-race/ethnicity marriage to shape different dimensions of mother-child relationships, including emotional ties to mothers and contact with mothers among young adults today. I focus on mothers because mother-child relationships tend to be more enduring than ties to fathers given women’s roles as kin-keepers (Hagestad, 1986; Rosenthal, 1985) as well as the absence of fathers for many young adults raised in single-mother households (Pew Research Center, 2015). I account for ties to mothers prior to marriage, during adolescence, which may select individuals into choosing same-race or different race partners. In addition, because previous research suggests that marriage outside of one’s race/ethnicity may be more detrimental for individuals of certain groups (e.g., Whites) compared to others (Kasinitz et al., 2008; Lee & Bean 2010; Lewis & Yancey, 1995; Root, 2001; Rosenblatt et al., 1995), I also assess whether the effects of intermarriage vary by offspring race/ethnicity and gender.

1. Background

1.1. Intermarriage and ties to parents

Previous research largely glosses over partner traits when considering the effect that marriage has on an individual’s ties to the family of origin. Whereas the majority of evidence finds that marriage has little effect on emotional ties to parents (Musick & Bumpass, 2012; Sarkisian & Gerstel, 2008; Spitze et al., 1994), married offspring visit less and have less contact with parents than those who are single (Bucx et al., 2008; Treas & Gubernskaya &, 2012; Kalmijn & DeVries 2009; Sarkisian & Gerstel, 2008; Spitze et al., 1994; Waite & Harrison, 1992). In addition, these weaker ties to parents fail to “bounce back” after an initial newlywed period (Musick & Bumpass, 2012).

Yet intermarriage may present a unique set of challenges to maintaining intergenerational ties. Prior studies find that parental and family opposition to interracial/ethnic dating and intermarriage remains common, despite public opinion in favor of intermarriage increasing over time. Interviews with college students, for example, suggest that parents’ disapproval of interracial/ethnic relationships underscores students’ decisions to not date across racial/ethnic boundaries and also deters the shift from dating to more committed relationships such as marriage (Childs, 2005; Harris & Kalbfleish, 2000). Interviews with and small surveys of intermarried couples cite disapproving reactions from family, friends, and acquaintances before and after marriage (Lewis & Yancey, 1995; Nemoto, 2009; Root, 2001). Family opposition in turn leads couples to limit their contact with kin (Killian, 2013; McNamara et al., 1999), in some instances because parents disowned offspring (Rosenblatt et al., 1995). Intermarried individuals also note the emotional stress of handling ties to disapproving family members throughout the relationship (Childs, 2005). A growing body of research finds that individuals who are intermarried are more prone to marital stress and mental stress than those in same-race/ethnicity marriages (Bratter & Eschbach, 2006; Fusco, 2010; Kroeger & Williams, 2011) and that this mental health disparity also exists in dating relationships (Tillman & Miller, 2017). These studies lend some support to the large body of qualitative findings that interracial/ethnic couples face challenging familial barriers and hint at the ways in which intermarried couples may have less social and familial support than those in same-race/ethnic marriages.

An important point, however, is that offspring who eventually intermarry may also have weaker ties to parents even before marriage. Evidence from research on adolescents finds that teens in interracial/ethnic dating relationships tend to report more distant relationships to parents than those in same-race relationships (Tillman & Miller, 2017; Wang, Kao & Joyner, 2006). Those who more frequently talk to parents are also less likely to report being in interracial and interethnic dating relationships (Huijnk & Liefbroer, 2012; Wang et al., 2006). In addition, adolescents who date across racial/ethnic lines meet their partner’s parents less frequently than those with same-race partners (Vaquera & Kao, 2005). These findings highlight how children with close ties to parents may find it difficult to introduce partners who are not readily integrated into tight-knit families. Thus, accounting for ties to parents prior to marriage is an important step in assessing whether intermarriage itself weakens ties to parents.

1.2. The role of race/ethnicity and gender

Research on intergenerational ties during adolescence and young adulthood points to substantial racial/ethnic differences. Scholars have noted the role of familismo in Hispanic families (Landale & Oropesa, 2007), and filial piety in East Asian families (Zhou & Bankston, 1998) as cultural factors that distinguish parent-child relationships in these groups compared to Whites. In fact, cultural factors as well as more constrained socioeconomic conditions likely reinforce one another to explain the higher share of intergenerational coresidence (Britton, 2013; Lei & South, 2016) among Blacks and Hispanics, for example, that are strongly associated with instrumental support to parents (Sarkisian & Gerstel, 2004; Sarkisian, Gerena, & Gerstel, 2007). Prior work on emotional closeness between generations provides mixed results. Whereas some studies report stronger emotional ties among non-White groups such as Hispanics, other studies show little to no difference across groups (see Landale & Oropesa, 2007 for a review). Furthermore, the effect of partner race on maintaining ties to parents may matter more for certain individuals compared to others. Much of the previous research on intermarriage takes note of differences in family members’ reactions to partner race depending on the family’s own racial/ethnic background as well as the gender of the person involved. One recent poll found that 72% of Blacks would accept a family member marrying someone of a different race/ethnicity, compared to only 64% of Asians, 63% of Hispanics and 61% of Whites (Wang, 2012). These findings support data from interviews with Black-White and Hispanic-White couples, where White partners more frequently cite strong family opposition to intermarriage than Black or Hispanic partners (Lewis & Yancey, 1995; Killian, 2013; McNamara et al., 1999; Root, 2001; Rosenblatt et al., 1995).

Gender may also influence the link between partner choice and ties to parents. On the one hand, as women tend to be kin-keepers, whom daughters marry may have less impact on their relationships with parents compared to sons. In studies comparing married to never married individuals, the gap between daughters’ contact with parents is smaller than the gap for sons, suggesting that marriage is less detrimental for daughters’ ties to parents than sons’ relationships (Sarkisian & Gerstel, 2004; 2008). A daughter’s ties to her parents may simply be more resilient over the life course (Merrill, 2011) and therefore – the strength of these relationships may be less sensitive to whom she marries. On the other hand, daughters’ romantic relationships may be subject to greater scrutiny than sons’ precisely because they are expected to be kin-keepers. In general, studies find that men report less disapproval from parents over interracial/ethnic dating than women (Park, Vo, & Tsung, 2009; Shenhav, Campos, & Goldberg, 2016). To summarize, it is possible that parents’ reactions to partner choice may differ by offspring gender.

2. The current study

Building upon prior research, this paper asks two questions. First, among married young adults today, do relationships to mothers differ among those who are intermarried compared to those in same-race/ethnicity marriages? To date, no large-scale studies using nationally representative data have explored this topic. The second research question asks whether the association between intermarriage and mother-child relationships varies depending on the race/ethnicity or gender of offspring. Previous findings from interviews with young adults suggest that the effects of intermarriage on intergenerational relationships, if any, likely differ according to these characteristics.

3. Method

3.1. Data and sample

This project uses data from the National Longitudinal Study of Adolescent to Adult Health (Add Health), a survey of adolescents in grades 7-12 in the United States during the 1994-95 school year. Approximately 20,745 adolescents were interviewed at Wave 1, when the average respondent was aged 16, with ages ranging from 12 to 21. The adolescents were followed into young adulthood with three additional in-home interviews in 1996, 2002 and in 2008/9. Of the 19,560 Wave 1 participants who were alive, in the country, and not on active military duty in Wave 4, the response rate was 80.27%, yielding 15,701 of the original respondents (Brownstein et al., 2011), when the average respondent was age 29. Crucial for this study is that a complete roster of all romantic relationships that the respondent ever had was collected at Wave 4. The respondent was asked to designate the start and end of these relationships, whether the relationship was still current, and the nature of the relationship (e.g., dating, cohabiting, or married, etc.). In addition, information was collected on the gender, race/ethnicity, and age of each partner.

The analytical sample uses information from respondents who were interviewed at Wave 1 and re-interviewed at Wave 4 only. The subsequent analysis focuses on how interracial/ethnic marriage affects ties to biological mothers in the United States. As kin-keepers, mothers maintain closer ties to offspring than fathers over the life course (Hagestad, 1986; Rosenthal, 1985). In addition, many children in the U.S. today are raised in single-mother families with little to no contact with non-residential biological fathers (Moffitt & Rendall, 1995). In the Add Health sample, approximately 42% of respondents reported living apart from biological fathers at Wave 1.

To create the analytical sample, I first excluded individuals who were not living with a biological mother at Wave 1 or Wave 2 (n=1,632). Next, I removed respondents who were already married at Wave 1 (n=46) when the respondents were still adolescents. I then excluded any respondents not married at Wave 4 (n= 8,198) in order to focus the sample on currently married young adults. Individuals who identified solely as Native American, Other, or were mixed race but most closely identified as Native American or Other were dropped from the analytical sample because of small sample sizes (n=79). Respondents whose biological mothers had died (n=293) or who did not report their biological mother as the main “mother figure” in Wave 4 were also excluded (n=185). Sibling (half-sibling, twin) pairs and respondents in the genetic oversample were not assigned weights and in the analysis that follows, I exclude all individuals without weights at Wave 1 in order to increase the representativeness of the data (n=237) (Chantala, 2006). Individuals with missing values on any of the outcome measures (n=6) were likewise removed from the sample.

Note that respondents who cohabited with their partners at Wave 4, but were not married, were also removed from the analytic sample. Although cohabiting individuals are an important population to examine, cohabitation is an informal institution that differs from marriage (Batson, Qian, & Lichter, 2006; Seltzer, 2000; Smock, 2000), which through its very public nature confers legal structures on the couple that differs from other relationships (Cherlin, 2004). Marriage establishes recognized norms of interacting with in-laws that are less well-defined for cohabiting couples. These expectations are important for married couples when negotiating competing obligations to the family of origin. In addition, prior research finds that the social barriers to interracial dating and cohabitation are much lower than marriage (Herman & Campbell, 2012; Zantvliet, Kalmijn & Verbakel, 2012). Thus, I only consider currently married respondents in this analysis but underscore the importance of examining differences between those who cohabit and those who marry for future research.

These exclusions produce an analytical sample of 5,025 individuals, which I use for the analysis of emotional closeness. I use a smaller sample of 4,712 respondents to examine visits and contact with mothers, given that coresident offspring are not asked to report on questions regarding maternal contact. Furthermore, less than 5% of the sample respondents had one or more missing values on the explanatory variables. To address this, I use the -mi- command suite in Stata 13.1 and create five multiply imputed datasets for the analysis.

3.2. Measures

I use three measures that capture different dimensions of intergenerational relationships during young adulthood: 1) the degree to which young adults have positive feelings about their mothers, 2) in-person visits with mothers, and 3) phone/email communication with mothers. In the Add Health survey, respondents are asked to report on how close they feel to their mothers at Wave 4, and answers range from not at all (1) to very much (5). In addition, two measures are used to capture mother-child contact. The first measure asks how often respondents visit their mother. The second measure assesses how often children communicate via phone, emails or letters with mothers. Answers range from never (0) to almost daily (5). Respondents who currently reside with their mother are not asked these questions and are excluded from the analyses that estimate these outcome measures.

Individuals are assigned values depending on whether they are intermarried or married to same race/ethnicity partners. Individuals who marry within the same race/ethnic group are defined as those who marry within the broad race/ethnic categories of non-Hispanic White, non-Hispanic Black, Asian, and Hispanic. Mixed race individuals were asked to report the race/ethnicity with which they most closely identified and were classified in that group. In separate analyses (not shown here), mixed race individuals were dropped from the analysis to assess the robustness of results (n=137). The substantive results remain similar without mixed race respondents. Respondents were asked to identify only one race/ethnicity for their partner and a mixed-race category was not provided as an option.

I include several control variables in the models. Respondent’s immigrant status is assigned to two categories: foreign-born respondents and native-born respondents with at least one foreign-born parent are categorized as children of immigrants, whereas native-born respondents with both native-born parents are classified as children of natives. I also include the respondent’s age, whether the respondent was a parent at Wave 4, the respondent’s educational attainment at Wave 4 (no high school, high school, some college, and college or more) and labor force status at Wave 4 (not employed, employed less than 35 hours, or 35 or more hours per week). In addition to the respondent’s characteristics, I also consider mothers’ and household traits at Wave 1. I include mothers’ educational attainment (less than high school, high school/GED, some college, college or more), and the respondent’s family type at Wave 1 (living with biological parents, a biological mother and a stepfather, or a biological mother only).

Finally, I also account for mother-child relationships during the respondent’s adolescence through a measure of emotional closeness to mothers, interaction with mothers, and communication with mothers. Emotional closeness to mothers at Wave 1 is the same question that is asked at Wave 4 and ranges from not at all (1) to very much (5). Interaction with mothers is captured with a combined scale where respondents reported on whether they did the following with mothers in past week: shopping, playing a sport, going to church, going to a movie/museum/other cultural event, or working on a school project. These items are added together and ranged from no activities (0) to all activities (5). In addition, respondents are asked to report whether they talked to their mothers in the past four weeks about a personal problem they were having (1 = yes, 0 = no). All of these controls for relationships to parents at Wave 1 are included in the models that predict contact and visit with mothers, but the indicator for emotional closeness at Wave 1 is removed in models that predict emotional closeness to mothers at Wave 4. Emotional closeness during adolescence is strongly correlated with emotional closeness to parents during young adulthood, and its inclusion could bias the model coefficients by violating assumptions about the error term (Markus, 1979). Although geographic proximity to parents is available in the data, I choose not to include it in the models due to collinearity with visits and contact with parents.

3.3. Analytic Strategy

In this analysis, I assess each measure of intergenerational relationships with mothers separately: emotional closeness, in-person visits, and phone/email contact. All analyses are performed using the -mi-, -svy- and -subpop- commands to account for the multiply imputed data and complex sampling design in Add Health. Data are weighted to adjust for sample selection and participation, as well as oversampling (Chantala & Tabor, 1999).

I begin by describing the sample of married respondents in Table 1. Next, step-wise multinomial logistic regression models are presented in Table 2 to assess how intermarriage affects emotional closeness to mothers at Wave 4. Intermarriage is first assessed (Model 1), followed by offspring demographics, mother and household characteristics (Model 2), and relationships to mothers prior to marriage are entered into the final model (Model 3). I repeat the same sequence of models for all outcomes and a summary of the relevant results for intermarriage is presented in Table 3 but the full results of the multinomial models are presented in online Appendix Tables A1 and A2. Next, I test for the significance of interactions between intermarriage and the respondent’s race/ethnicity and gender. I do not present the full models from these interactions, but present a summary of the results for moderation by race/ethnicity in Table 4 and gender in Table 5. The full models from the results are provided in online Appendix Tables A3 to A8.

Table 1. Respondent Characteristics (N= 5,025).

	Proportion/ Mean	SE		Proportion/ Mean	SE
Respondent's traits at W4			Relationship to mother at W1
Marriage type			Emotional closeness	0.68	0.01
Intermarriage	0.14		Interaction1	1.65	0.03
Same-race marriage	0.86		Talk - personal problem	0.40
Race/ethnicity			Relationship to mother at W4
NH White	0.77		Emotional closeness
NH Black	0.08		Not at all	0.01
NH Asian	0.03		Very little	0.02
Hispanic (all races)	0.12		Somewhat	0.09
Child of immigrant(s)	0.15		Quite a bit	0.20
Female	0.54		Very much	0.68
Age	29.20	0.11	Visit2
Parent	0.66		Never	0.01
Education			<=1 year	0.05
Less than high school	0.06		Few/year	0.19
High school/GED	0.15		1/2 month	0.23
Some college	0.44		1/2 Week	0.33
College or more	0.35		Almost daily	0.19
Labor force status			Phone/Email contact2
Not employed	0.18		Never	0.01
Part-time	0.10		<=1 year	0.01
Full-time	0.71		Few/year	0.01
Mother and Household's traits			1/2 month	0.08
Mother's education at W1			1/2 Week	0.40
<8grade	0.05		Almost daily	0.48
8<grade<12	0.11
HS/GED	0.37
Some college	0.21
College or more	0.26
Family type at W1
Bio mother & bio father	0.69
Bio mother & stepfather	0.09
Bio mother only	0.21

Note: All descriptives are weighted and adjusted for AddHealth's complex survey design.

¹Respondents reported on whether they did the following with mothers in past week: shopping, playing a sport, going to church, going to a movie/museum/other cultural event, worked on a school project. These items were added together (0=no activities, 5=all activities).

²Excludes Rs who live with mothers. N= 4,712

Table 2. Parameters from Logistic Regression Models that Predict Being "Very" Emotionally Close to Mothers (N= 5,025).

	Model 1		Model 2		Model 3
	OR	SE B	OR	SE B	OR	SE B
Omitted: Less tha "very" close ties
Respondent's traits
Intermarriage	0.87	0.12	0.81	0.13	0.81	0.13
Race/ethnicity (/NH White)
NH Black			2.85	0.20 ***	2.83	0.21 ***
NH Asian			0.80	0.33	0.82	0.33
Hispanic (all races)			1.65	0.17 **	1.66	0.18 **
Child of immigrants			0.78	0.18	0.78	0.18
Female			1.54	0.09 ***	1.44	0.10 ***
Age			0.96	0.03	0.97	0.03
Parent			0.75	0.09 **	0.75	0.09 **
Education (/< High school)
High school/GED			1.01	0.22	0.98	0.22
Some college			1.06	0.19	1.00	0.20
College or more			1.15	0.22	1.04	0.22
Labor force status (/not employed)
Part-time			1.14	0.17	1.11	0.17
Full-time			1.16	0.12	1.14	0.12
Mother's and household's traits
Mother's education at W1 (/< 8 grade)
8<grade<12			0.81	0.24	0.81	0.25
High school/GED			0.84	0.24	0.81	0.24
Some college			0.80	0.24	0.75	0.24
College or more			0.70	0.24	0.67	0.24
Family type at W1 (/both bio parents married)
Bio mother and stepfather			0.73	0.14 *	0.73	0.14 *
Bio mother only			0.74	0.12 *	0.73	0.12 **
Relationship to mother at W1
Emotional closeness					1.15	0.04 **
Frequency of interaction					1.26	0.09 **
Talk about personal problem					4.12	0.98
Constant	2.16	0.05 ***	7.01	0.92 *	0.81	0.13

Note: Models are weighted and adjusted for AddHealth's complex survey design.

^***p<.001;

^**p<.01;

^*p<.05

Table 3. Summarized Parameters from Models Predicting Relationship to Mothers.

	Model 1		Model 2		Model 3
	OR	SE B	OR	SE B	OR	SE B
Panel A. Models predicting being “very” emotionally close to mothers (N=5,025)
Omitted: Less than "very" close ties
Intermarriage	0.87	0.12	0.81	0.13	0.81	0.13
Panel B. Models predicting visits with mothers (N=4,712)1
Omitted: Less than weekly
Weekly
Intermarriage	1.00	0.12	0.90	0.11	0.90	0.11
Daily
Intermarriage	0.95	0.13	0.88	0.14	0.88	0.14
Panel C. Models predicting phone/email contact with mothers (N=4,712)1
Omitted: Less than weekly
Weekly
Intermarriage	0.82	0.17	0.72	0.19	0.72	0.19
Daily
Intermarriage	0.86	0.19	0.75	0.22	0.76	0.22

Note:

¹Excludes Rs who coreside with mothers at W4. All models include the same control variables as in Table 2. Models are weighted and adjusted for AddHealth's complex survey design.

^***p<.001;

^**p<.01;

^*p<.05

Table 4. Summarized Parameters from Models Predicting Relationship to Mothers by Respondent Intermarriage and Race/ethnicity.

	OR	SE B
Panel A. Models predicting being “very” emotionally close to mothers (N=5,025)
Omitted: Less than "very" close ties
Intermarriage	0.94	0.18
Race/ethnicity (/NH White)
NH Black	3.02	0.23 ***
NH Asian	0.90	0.37
Hispanic (all races)	1.99	0.22 *
Intermarriage × Race/ethnicity
Intermarriage × NH Black	0.71	0.40
Intermarriage × NH Asian	0.75	0.57
Intermarriage × Hispanic	0.62	0.35
Panel B. Models predicting visits with mothers (N=4,712)1
Omitted: Less than weekly
weekly
Intermarriage	0.94	0.15
Race/ethnicity (/NH White)
NH Black	0.98	0.18
NH Asian	1.50	0.41
Hispanic (all races)	1.29	0.24
Intermarriage × Race/ethnicity
Intermarriage × NH Black	0.84	0.36
Intermarriage × NH Asian	0.34	0.54 *
Intermarriage × Hispanic	1.12	0.29
Daily
Intermarriage	1.10	0.18
Race/ethnicity (/NH White)
NH Black	1.81	0.14 ***
NH Asian	1.49	0.50
Hispanic (all races)	1.58	0.28
Intermarriage × Race/ethnicity
Intermarriage × NH Black	0.33	0.42 **
Intermarriage × NH Asian	0.59	0.66
Intermarriage × Hispanic	0.72	0.37
Panel C. Models predicting phone/email contact with mothers (N=4,712)1
Omitted: Less than weekly
Weekly
Intermarriage	0.79	0.26
Race/ethnicity (/NH White)
NH Black	1.72	0.29
NH Asian	1.97	0.50
Hispanic (all races)	2.41	0.46
Intermarriage × Race/ethnicity group
Intermarriage × NH Black	0.50	0.60
Intermarriage × NH Asian	0.70	0.75
Intermarriage × Hispanic	1.01	0.63
Daily
Intermarriage	0.85	0.28
Race/ethnicity (/NH White)
NH Black	3.11	0.29 ***
NH Asian	1.11	0.51
Hispanic (all races)	2.33	0.45
Intermarriage × Race/ethnicity
Intermarriage × NH Black	0.45	0.57
Intermarriage × NH Asian	0.64	0.76
Intermarriage × Hispanic	0.97	0.63

Note:

¹Excludes Rs who coreside with mothers at W4. All models include the same control variables as in Table 2. Models are weighted and adjusted for AddHealth's complex survey design.

^***p<.001;

^**p<.01;

^*p<.05

Table 5. Summarized Parameters from Models Predicting Relationship to Mothers by Respondent Intermarriage and Gender.

	OR	SE B
Panel A. Models predicting being “very” emotionally close to mothers (N= 5,025)
Omitted: Less than "very" close ties
Intermarriage	0.88	0.19
Female	1.70	0.11 ***
Female × Intermarriage	0.87	0.27
Panel B. Models predicting visits with mothers (N=4,712)1
Omitted: Less than weekly
Weekly
Intermarriage	0.91	0.15
Female	1.42	0.10 **
Female × Intermarriage	0.98	0.23
Daily
Intermarriage	0.83	0.21
Female	2.40	0.14 ***
Female × Intermarriage	1.09	0.33
Panel C. Models predicting phone/email contact with mothers (N=4,712)1
Omitted: Less than weekly
Weekly
Intermarriage	0.94	0.22
Female	2.65	0.20 ***
Female × Intermarriage	0.44	0.38 *
Daily
Intermarriage	1.01	0.25
Female	11.20	0.20 ***
Female × Intermarriage	0.45	0.41

Note:

¹Excludes Rs who coreside with mothers at W4. All models include the same control variables as in Table 2. Models are weighted and adjusted for AddHealth's complex survey design.

^***p<.001;

^**p<.01;

^*p<.05

Although the outcome measures lend themselves well to ordered response models, Brant tests (not shown here) indicate that the parallel regression assumption is violated for all outcome measures. In separate analyses, I also applied generalized ordered logit models (Williams, 2006). However, because many of the tests used to detect differences on subpopulations using imputed data could not be used with generalized ordered logit models in Stata, I use multinomial logistic regression models and combine categories so that the results are more interpretable. For each outcome measure, I re-categorize the responses into groups that theoretically capture the range of possible responses, although several variations of cut points and category groupings were originally explored. For emotional closeness, the categories are re-grouped so that individuals report being less than very close or very close to mothers. For measures of visits and contact, respondents can report visiting/talking with mothers less than weekly, weekly or daily. The base outcome category for all models is having the weakest relationship to mothers.

4. Results

Table 1 presents descriptive characteristics of the sample. Approximately 1 in 7 young adults in the sample – 14% – report being interracially married. This is comparable to 15% of all new marriages between people of different racial/ethnic backgrounds at the national level (Wang, 2012). Approximately 23% of respondents in the sample consist of non-Whites and 15% have at least one immigrant parent. Although the sample of married respondents is young, with an average age of 29, approximately two-thirds have already become parents by Wave 4. The vast majority – nearly 80% – have completed some college or more and over 70% work full-time. The majority of respondents in the sample grew up in two-parent biological households, with mothers having received less education, on average, than their children in the sample.

During adolescence (Wave 1), most respondents report being emotionally close to mothers, although mothers and children have low levels of interaction on a weekly basis. In addition, only 40% spoke with their mother about a personal problem in the last week. By young adulthood (Wave 4), respondents remain very emotionally close to mothers. Among those who did not live with their mothers, more than 50% reported visiting mothers on a weekly or daily basis, and over 80% reported having phone or email contact with mothers on at least a weekly basis. In general, these descriptive statistics suggest strong, positive, relationships with mothers among this sample of married young adults.

The next section addresses whether being intermarried or married to a same-race/ethnicity partner differentiates offspring relationships with mothers. Table 2 presents results from the nested models: Model 1 includes an indicator for whether or not the young adult is intermarried; the results suggest that there is no association between partner type and emotional closeness to mothers. Model 2 adds in children’s and mothers’ characteristics and the relationship between partner type and emotional closeness to mothers remains insignificant. However, the model results show that Black and Hispanic children report greater closeness to mothers than non-Hispanic Whites. Daughters also experience closer emotional ties to mothers than sons. However, those offspring raised in step-families or by single mothers report less closeness to mothers than those raised in two-parent biological families. This does not change with the addition of relationship to mothers during adolescence in Model 3. Results from Model 3 also confirm that individuals who are emotionally close to mothers as adolescents remain close to mothers as young adults. However, intermarriage remains insignificantly associated with closeness to mothers in young adulthood.

Table 3 presents summary results of the effect of intermarriage from similar multinomial logistic models predicting visits and contact with mothers during young adulthood. The results for emotional closeness to mothers (Table 2) are also presented in the first panel of Table 3 for comparison. Model progression is the same as reported in Table 2. In Panels B and C, it is evident that there are no significant differences in the frequency of visits and phone/email contact with mothers between adult children who are intermarried compared to those in same-race/ethnicity unions. These results echo those of emotional closeness (Panel A), suggested that intermarriage has little overall effect on children’s relationships with mothers among those who are currently married.

Finally, a series of additional analyses were conducted to assess whether respondent race/ethnicity or gender moderated the relationship between intermarriage and mother-child relationships during adulthood. A summary of the results for the moderating effect of children’s race/ethnicity is presented in Table 4 and the full models are presented in online Appendix Tables A3 – A5. In general, the interaction models show few significant effects, although one finding is worth mentioning. In Panel B, the results suggest that Asians experience the effects of intermarriage differently than Whites with respect to weekly visits, and Blacks also experience the effects of intermarriage differently than Whites with respect to daily visits. However, these results are difficult to interpret without accounting for the main effects of race/ethnicity.

To better interpret this interaction, and to include the main effects of race/ethnicity, Figure 1 presents the predicted probabilities of visiting mothers weekly and daily by offspring race/ethnicity and intermarriage status. In Figure 1A, there are no significant differences across racial/ethnic groups by partner status in the likelihood of visiting mothers on a weekly basis. Specifically, the confidence intervals for intermarried Asians and Asians married to other Asians overlap once the main effects of race/ethnicity are taken into account. However in Figure 1B, we see that among Blacks, intermarried respondents are half as likely (13%) to visit mothers daily, compared to Blacks married to black spouses (26%). Among other race/ethnic groups, contrasts between those who are intermarried and not intermarried are not significantly different.

Figure 1.

A: Predicted Probability of Weekly Visits

B: Predicted Probability of Daily Visits

Table 5 presents results for the moderation analysis by gender. Similar to results for the moderation analysis by race/ethnicity, few significant differences are apparent. Only in Panel C do model results indicate that daughters who marry outside of their racial/ethnic group are less likely to speak with mothers on a weekly daily basis compared to sons who are intermarried. However, Figure 2 presents the predicted probabilities to better interpret the interaction results. In Panel A, sons have a greater likelihood of contacting mothers on a weekly basis, compared to daughters, but Panel B demonstrates that daughters are more likely to be in phone/email contact with mothers on a daily basis, compared to sons. When looking at differences among sons or among daughters, however, no significant variation by partner type is apparent. Daughters who are married to same-race/ethnicity partners do not contact mothers more frequently than daughters who are intermarried and the same can be said of sons. The central finding from the moderation analysis is that the relationship between children’s partner type and ties to mothers is rarely dependent on offspring race/ethnicity or gender, with few significant differences within groups by intermarriage status.

Fig. 2.

A: Predicted Probability of Weekly Phone/Email Contact

B: Predicted Probability of Daily Phone/Email Contact

5. Discussion

Interviews with intermarried couples hint at the ways in which immediate and extended family members voice opposition both prior to and following marriage (Killian, 2013; McNamara et al., 1999; Rosenblatt et al., 1995). Even before marriage, research on adolescents finds that those who date interracially/ethnically communicate less with parents and are less open about their dating lives than those who date same-race partners (Huijnk & Liefbroer, 2012; Vaquera & Kao, 2005; Wang et al., 2006). This study, however, finds no evidence to support the idea that intermarriage in general is associated with weaker ties to mothers. In attempting to understand whether the effects of marriage varied by offspring race/ethnicity or gender, almost no significant effects were found. Despite a large breadth of evidence suggesting greater damage to parent-child relationships among Whites who intermarry compared to non-Whites (Lewis & Yancey, 1995; Root, 2001; Rosenblatt et al., 1995) findings from this study offer little support for this argument. In conducting a large number of tests, only one significant finding emerged: Blacks who intermarried tend to visit mothers less frequently than those with Black spouses. One explanation is that Blacks who intermarry are more open to living in, and perhaps more importantly, are accepted into neighborhoods where Blacks are not the majority. Previous research finds that Black-White couples, for example, are more likely to live in racially diverse neighborhoods than Whites or Blacks in same-race unions (Wright, Ellis, & Holloway, 2013). This geographic mobility among intermarried Blacks could decrease contact between young adults and their parents. Despite this, the main point remains the same, in that intermarriage has few differential effects on relationships to mothers by children’s race/ethnicity or gender.

However, this study is limited in a number of ways that could constrain the interpretability of its findings. First, this project examines marriages only and does not include other types of unions. Offspring in cohabiting unions tend to have less contact with parents than those who are married, although the difference often depends on the social context (Nazio & Saraceno, 2010; Yahirun and Hamplova, 2014). Also, cohabitation with a partner of a different racial/ethnic origin than one’s own may have fewer consequences for parent-child relationships if parents do not view these unions as permanent (Zantvliet et al., 2012). Thus, whereas the addition of cohabiting couples would add another layer of complexity to this study, I leave this area for future research.

Second, the data used here include relatively young individuals – adults aged 25 to 34 in 2008/9. These results are thus representative of marriages which occur relatively early on and may not be representative of intermarriages in general, which previous studies show tend to occur later in life (Lichter, 1990). In addition, qualitative data from interviews with young adults imply that parental resistance to intermarriage tends to decrease with age (Kasinitz et al., 2008). Additional waves of Add Health data collection should help address whether the effects of intermarriage remain the same across the life course.

Third, the analysis presented here does not distinguish between specific couple types. However, qualitative research shows that certain couples, such as Black-White couples, continue to face strong familial and societal disapproval (McNamara et al., 1999; Root, 2001; Rosenblatt et al., 1995). In addition, among intermarried couples, couples where one partner is White may differ from couples where both partners are non-White. In supplementary analyses, I further constrained the sample to intermarried couples and found that couples where both partners were racial/ethnic minorities had closer emotional ties to mothers than those couples with one White partner (results not shown). This pattern suggests that couple’s specific racial composition may matter for certain outcomes, but splitting the sample into specific racial/ethnic combinations is not possible in the Add Health data due to its small sample of intermarriages. Future data collection in national surveys would enable a more detailed examination of how one’s own race and partner’s race matter for ties to the family of origin.

Fourth, this paper only examines spouses’ race/ethnicity and how that influences ties to mothers. It is likely that partner’s educational background and socioeconomic status, for example, are equally important dimensions that also influence an individual’s ties to parents. However, the Add Health survey does not collect this information. In addition, the analysis presented here does not include ties to biological fathers, and focuses on biological mothers only because of the large number of respondents with non-resident biological fathers at Wave 1. Future work should address how intermarriage affects ties to fathers as well.

Furthermore, the link between intermarriage and mother-child ties could be influenced by a myriad of factors other than the young adult’s race/ethnicity or gender. Recent research, for example, suggests that intermarried couples with children report less support from extended kin than same-race couples with children (Bratter and Whitehead, 2018). In supplemental analyses, however, I found that having a child did not significantly moderate the link between intermarriage and mother-child relationships in young adulthood.

Finally, the results presented here pay little attention to the sequencing of life events that lead to intermarriage that originate with ties to parents. Children with close relationships to mothers as adolescents may choose to remain nearby as young adults. This likely shapes their choices on the marriage market, if we assume that partner choices are geographically constrained. Because Americans tend to live in racially and ethnically segregated neighborhoods (Wright et al., 2013), those who continue to live close to parents are thus more likely to choose same-race/ethnicity partners. Although this study controls for relationships to mothers prior to marriage, a separate study examining whether children’s relationships to parents during adolescence lead to specific union formation outcomes is needed.

Nonetheless, findings from this study contribute to a broader understanding of how marriage and intermarriage in particular influence ties between offspring and their mothers for a generation of young adults who are entering into relationships 25 years after the U.S. Supreme Court’s legalization of interracial marriage. Despite the increasing acceptability of intermarriage as expressed in national opinion polls (Newport, 2013), new marriages that consist of couples of different racial/ethnic origins remain relatively uncommon (Wang, 2012). Whereas prior research hints at family conflict and opposition as reasons why intermarriage rates remain low (Lewis & Yancey, 1995; Killian, 2013; McNamara et al., 1999; Root, 2001; Rosenblatt et al., 1995), findings from this study suggest that once married, children in intermarried couples tend to have very similar relationships with mothers compared to those who marry same-race/ethnicity partners. It is possible that intermarriage today remains both an indicator and a mechanism for racial and ethnic integration in diverse social contexts such as the United States (Gordon, 1964), if ties to the family of origin remain as strong among intermarried couples as they do among those who married within racial/ethnic boundaries. However, these results should be paired with data documenting the decline of marriage overall and the greater propensity for young adults today to delay marriage, or cohabit both before or in lieu of marriage (Kennedy & Bumpass, 2008; Smock 2000). More work is needed to understand the effect of partner characteristics on ties to the family of origin beyond the context of marriage. In addition, as the United States becomes increasingly racially and ethnically diverse, future research should assess how ties to parents during adolescence shape decisions to choose partners that cross racial/ethnic boundaries in adulthood.

Appendix

Table A1.

Parameters From Multinomial Logistic Regression Models Predicting Visits to Mothers by Partner Type (N= 4,712)

		M1		M2		M3
	OR	SE B	OR	SE B	OR	SE B
Omitted: Less than weekly
weekly
Respondent's traits
Intermarriage	1.00	0.12	0.90	0.11	0.90	0.11
Race/ethnicity (/NH White)
NH Black			0.98	0.17	0.95	0.17
NH Asian			1.09	0.37	1.11	0.37
Hispanic (all races)			1.40	0.21	1.38	0.21
Child of immigrants			1.23	0.17	1.23	0.17
Female			1.35	0.10 **	1.42	0.10 ***
Age			0.96	0.03	0.97	0.03
Parent			1.17	0.11	1.18	0.11
Education (/< High school)
High school/GED			0.69	0.23	0.68	0.23
Some college			0.65	0.21 *	0.66	0.22
College or more			0.44	0.22 ***	0.45	0.23 **
Labor force status (/not employed)
Part-time			1.62	0.16 **	1.64	0.16 **
Full-time			1.44	0.13 **	1.44	0.13 **
Mother's and household's traits
Mother's education at W1 (/< 8 grade)
8<grade<12			0.78	0.28	0.77	0.27
High school/GED			0.94	0.24	0.93	0.24
Some college			0.82	0.28	0.83	0.28
College or more			0.63	0.26	0.63	0.26
Family type at W1 (/both bio parents married)
Bio mother and stepfather			0.59	0.16 **	0.59	0.16 **
Bio mother only			0.66	0.10 ***	0.66	0.10 ***
Relationship to mother at W1
Emotional closeness					1.27	0.09 *
Frequency of interaction					0.96	0.05
Talk about personal problem					0.88	0.08
Constant	0.70	0.06 ***	2.96	0.92	2.19	0.93
Daily
Respondent's traits
Intermarriage	0.95	0.13	0.88	0.14	0.88	0.14
Race/ethnicity (/NH White)
NH Black			1.63	0.15 **	1.54	0.14 **
NH Asian			1.25	0.42	1.28	0.41
Hispanic (all races)			1.52	0.23	1.48	0.23
Child of immigrants			0.77	0.22	0.77	0.23
Female			2.26	0.13 ***	2.43	0.13 ***
Age			0.99	0.04	1.01	0.04
Parent			2.24	0.15 ***	2.28	0.15 ***
Education (/< High school)
High school/GED			0.70	0.28	0.68	0.28
Some college			0.53	0.25 *	0.52	0.26 *
College or more			0.24	0.27 ***	0.22	0.28 ***
Labor force status (/not employed)
Part-time			1.54	0.18 *	1.56	0.18 *
Full-time			1.53	0.18 *	1.51	0.18 *
Mother's and household's traits
Mother's education at W1 (/< 8 grade)
8<grade<12			0.62	0.32	0.61	0.31
High school/GED			0.58	0.31	0.56	0.30
Some college			0.46	0.33 *	0.46	0.33 *
College or more			0.29	0.33 ***	0.29	0.32 ***
Family type at W1 (/both bio parents married)
Bio mother and stepfather			0.52	0.17 ***	0.52	0.17 ***
Bio mother only			0.57	0.15 ***	0.58	0.15 ***
Relationship to mother at W1
Emotional closeness					1.56	0.13 **
Frequency of interaction					1.04	0.06
Talk about personal problem					0.85	0.11
Constant	0.41	0.08 ***	0.78	1.13	0.27	1.12

Note: Models are weighted and adjusted for AddHealth's complex survey design.

^***p<.001;

^**p<.01;

^*p<.05

Table A2.

Parameters From Multinomial Logistic Regression Models Predicting Phone/Email Contact with Mothers (N= 4,712)

		M1		M2		M3
	OR	SE B	OR	SE B	OR	SE B
Omitted: Less than weekly
Weekly
Respondent's traits
Intermarriage	0.82	0.17	0.72	0.19	0.72	0.19
Race/ethnicity (/NH White)
NH Black			1.53	0.25	1.40	0.25
NH Asian			1.75	0.37	1.78	0.37
Hispanic (all races)			2.56	0.33 **	2.50	0.32 **
Child of immigrants			0.82	0.28	0.81	0.27
Female			2.20	0.16 ***	2.31	0.17 ***
Age			0.99	0.04	1.02	0.05
Parent			0.77	0.13	0.78	0.13
Education (/< High school)
High school/GED			1.27	0.29	1.21	0.30
Some college			1.78	0.31	1.72	0.32
College or more			2.08	0.32 *	1.94	0.32 *
Labor force status (/not employed)
Part-time			0.69	0.28	0.68	0.28
Full-time			1.50	0.21	1.46	0.21
Mother's and household's traits
Mother's education at W1 (/< 8 grade)
8<grade<12			1.12	0.45	1.09	0.46
High school/GED			0.92	0.43	0.87	0.43
Some college			0.85	0.45	0.83	0.45
College or more			1.25	0.43	1.19	0.43
Family type at W1 (/both bio parents married)
Bio mother and stepfather			0.81	0.23	0.80	0.23
Bio mother only			0.61	0.15 **	0.62	0.15 **
Relationship to mother at W1
Emotional closeness					1.43	0.15 *
Frequency of interaction					1.11	0.07
Talk about personal problem					0.92	0.14
Constant	3.58	0.08 ***	1.80	1.34	0.58	1.53
Daily
Respondent's traits
Intermarriage	0.86	0.19	0.75	0.22	0.76	0.22
Race/ethnicity (/NH White)
NH Black			2.90	0.28***	2.49	0.27**
NH Asian			0.91	0.42	0.97	0.42
Hispanic (all races)			2.54	0.31 **	2.40	0.31 **
Child of immigrants			0.72	0.24	0.71	0.25
Female			8.66	0.16 ***	9.75	0.17 ***
Age			0.94	0.04	1.00	0.05
Parent			1.05	0.14	1.09	0.14
Education (/< High school)
High school/GED			0.95	0.33	0.87	0.35
Some college			1.26	0.30	1.18	0.32
College or more			0.96	0.30	0.83	0.32
Labor force status (/not employed)
Part-time			0.57	0.27 *	0.55	0.27 *
Full-time			1.32	0.20	1.25	0.21
Mother's and household's traits
Mother's education at W1 (/< 8 grade)
8<grade<12			1.34	0.46	1.28	0.48
High school/GED			1.02	0.39	0.93	0.40
Some college			0.91	0.40	0.86	0.41
College or more			0.91	0.40	0.84	0.41
Family type at W1 (/both bio parents married)
Bio mother and stepfather			0.62	0.26	0.60	0.26
Bio mother only			0.53	0.18 **	0.53	0.18 **
Relationship to mother at W1
Emotional closeness					2.69	0.18 ***
Frequency of interaction					1.19	0.07 *
Talk about personal problem					0.94	0.14
Constant	4.23	0.08 ***	7.75	1.35	0.63	1.63

Note: Models are weighted and adjusted for AddHealth's complex survey design.

^***p<.001;

^**p<.01;

^*p<.05

Table A3.

Parameters from Logistic Regression Models that Predict Being "Very" Emotionally Close to Mothers by Respondent Intermarriage and Race/ethnicity (N= 5,025)

	OR	SE B
Omitted: Less than "very" close ties
Respondent's traits
Intermarriage	0.94	0.18
Race/ethnicity (/NH White)
NH Black	3.02	0.23 ***
NH Asian	0.90	0.37
Hispanic (all races)	1.99	0.22 *
Intermarriage × Race/ethnicity
Intermarriage × NH Black	0.71	0.40
Intermarriage × NH Asian	0.75	0.57
Intermarriage × Hispanic	0.62	0.35
Child of immigrants	0.75	0.18
Female	1.43	0.10 ***
Age	0.97	0.03
Parent	0.75	0.09 **
Education (/< High school)
High school/GED	0.99	0.22
Some college	1.01	0.20
College or more	1.06	0.21
Labor force status (/not employed)
Part-time	1.11	0.17
Full-time	1.13	0.12
Mother's and household's traits
Mother's education at W1 (/< 8 grade)
8<grade<12	0.84	0.25
High school/GED	0.86	0.25
Some college	0.79	0.25
College or more	0.70	0.25
Family type at W1 (/both bio parents married)
Bio mother and stepfather	0.73	0.14 *
Bio mother only	0.73	0.12 **
Relationship to mother at W1
Frequency of interaction	1.15	0.04 **
Talk about personal problem	1.26	0.09 **
Constant	3.93	0.98

Note: Models are weighted and adjusted for AddHealth's complex survey design.

^***p<.001;

^**p<.01;

^*p<.05

Table A4.

Parameters From Multinomial Logistic Regression Models Predicting Visits to Mothers by Respondent Intermarriage and Race/ethnicity (N= 4,712)

	OR	SE B
Omitted: Less than weekly
Weekly
Respondent's traits
Intermarriage	0.94	0.15
Race/ethnicity (/NH White)
NH Black	0.98	0.18
NH Asian	1.50	0.41
Hispanic (all races)	1.29	0.24
Intermarriage × Race/ethnicity
Intermarriage × NH Black	0.84	0.36
Intermarriage × NH Asian	0.34	0.54 *
Intermarriage × Hispanic	1.12	0.29
Child of immigrants	1.25	0.17
Female	1.42	0.10 ***
Age	0.97	0.03
Parent	1.18	0.11
Education (/< High school)
High school/GED	0.68	0.23
Some college	0.66	0.22
College or more	0.44	0.23 **
Labor force status (/not employed)
Part-time	1.63	0.16 **
Full-time	1.43	0.13 **
Mother's and household's traits
Mother's education at W1 (/< 8 grade)
8<grade<12	0.75	0.28
High school/GED	0.91	0.24
Some college	0.82	0.29
College or more	0.62	0.27
Family type at W1 (/both bio parents married)
Bio mother and stepfather	0.59	0.16 **
Bio mother only	0.66	0.10 ***
Relationship to mother at W1
Emotional closeness	1.27	0.09 *
Frequency of interaction	0.96	0.05
Talk about personal problem	0.88	0.08
Constant	2.24	0.92
Daily
Respondent’s traits
Intermarriage	1.10	0.18
Race/ethnicity (/NH White)
NH Black	1.81	0.14 ***
NH Asian	1.49	0.50
Hispanic (all races)	1.58	0.28
Intermarriage × Race/ethnicity
Intermarriage × NH Black	0.33	0.42 **
Intermarriage × NH Asian	0.59	0.66
Intermarriage × Hispanic	0.72	0.37
Child of immigrants	0.76	0.23
Female	2.38	0.13 ***
Age	1.01	0.04
Parent	2.28	0.15 ***
Education (/< High school)
High school/GED	0.69	0.28
Some college	0.52	0.26 *
College or more	0.23	0.28 ***
Labor force status (/not employed)
Part-time	1.54	0.18 *
Full-time	1.49	0.18 *
Mother's and household's traits
Mother's education at W1 (/< 8 grade)
8<grade<12	0.62	0.31
High school/GED	0.57	0.31
Some college	0.47	0.33 *
College or more	0.29	0.33 ***
Family type at W1 (/both bio parents married)
Bio mother and stepfather	0.52	0.17 ***
Bio mother only	0.57	0.15 ***
Relationship to mother at W1
Emotional closeness	1.57	0.13 **
Frequency of interaction	1.04	0.06
Talk about personal problem	0.86	0.11
Constant	0.26	1.13

Note: Models are weighted and adjusted for AddHealth's complex survey design.

^***p<.001;

^**p<.01;

^*p<.05

Table A5.

Parameters From Multinomial Logistic Regression Models Predicting Phone/Email Contact with Mothers by Respondent Intermarriage and Race/ethnicity (N= 4,712)

	OR	SE B
Omitted: Less than weekly
Weekly
Respondent's traits
Intermarriage	0.79	0.26
Race/ethnicity (/NH White)
NH Black	1.72	0.29
NH Asian	1.97	0.50
Hispanic (all races)	2.41	0.46
Intermarriage × Race/ethnicity
Intermarriage × NH Black	0.50	0.60
Intermarriage × NH Asian	0.70	0.75
Intermarriage × Hispanic	1.01	0.63
Child of immigrants	0.82	0.28
Female	2.29	0.18 ***
Age	1.03	0.05
Parent	0.79	0.13
Education (/< High school)
High school/GED	1.21	0.30
Some college	1.72	0.32
College or more	1.94	0.32 *
Labor force status (/not employed)
Part-time	0.68	0.28
Full-time	1.44	0.21
Mother's and household's traits
Mother's education at W1 (/< 8 grade)
8<grade<12	1.07	0.46
High school/GED	0.86	0.43
Some college	0.82	0.46
College or more	1.18	0.43
Family type at W1 (/both bio parents married)
Bio mother and stepfather	0.81	0.23
Bio mother only	0.61	0.15 **
Relationship to mother at W1
Emotional closeness	1.44	0.15 *
Frequency of interaction	1.11	0.07
Talk about personal problem	0.92	0.14
Constant	0.57	1.54
Daily
Respondent's traits
Intermarriage	0.85	0.28
Race/ethnicity (/NH White)
NH Black	3.11	0.29 ***
NH Asian	1.11	0.51
Hispanic (all races)	2.33	0.45
Intermarriage × Race/ethnicity
Intermarriage × NH Black	0.45	0.57
Intermarriage × NH Asian	0.64	0.76
Intermarriage × Hispanic	0.97	0.63
Child of immigrants	0.72	0.26
Female	9.67	0.17 ***
Age	1.00	0.05
Parent	1.10	0.14
Education (/< High school)
High school/GED	0.87	0.34
Some college	1.19	0.32
College or more	0.83	0.32
Labor force status (/not employed)
Part-time	0.54	0.27 *
Full-time	1.23	0.21
Mother's and household's traits
Mother's education at W1 (/< 8 grade)
8<grade<12	1.25	0.48
High school/GED	0.92	0.40
Some college	0.85	0.41
College or more	0.83	0.41
Family type at W1 (/both bio parents married)
Bio mother and stepfather	0.61	0.26
Bio mother only	0.53	0.18 **
Relationship to mother at W1
Emotional closeness	2.70	0.18 ***
Frequency of interaction	1.19	0.07 *
Talk about personal problem	0.95	0.14
Constant	0.62	1.64

Note: Models are weighted and adjusted for AddHealth's complex survey design.

^***p<.001;

^**p<.01;

^*p<.05

Table A6.

Parameters from Logistic Regression Models that Predict Being "Very" Emotionally Close to Mothers by Respondent Intermarriage and Gender (N= 5,025)

	OR	SE B
Omitted: Less than "very" close ties
Respondent's traits
Intermarriage	0.83	0.18
Female	1.45	0.11 **
Intermarriage × Female	0.96	0.25
Race/ethnicity (/NH White)
NH Black	0.78	0.18
NH Asian	2.82	0.21 ***
Hispanic (all races)	0.82	0.33
Child of immigrants	1.66	0.18 **
Age	0.97	0.03
Parent	0.75	0.09 **
Education (/< High school)
High school/GED	0.98	0.22
Some college	1.00	0.20
College or more	1.04	0.22
Labor force status (/not employed)
Part-time	1.11	0.17
Full-time	1.14	0.12
Mother's and household's traits
Mother's education at W1 (/< 8 grade)
8<grade<12	0.81	0.25
High school/GED	0.81	0.24
Some college	0.75	0.24
College or more	0.67	0.24
Family type at W1 (/both bio parents married)
Bio mother and stepfather	0.73	0.14 *
Bio mother only	0.73	0.12 **
Relationship to mother at W1
Frequency of interaction	1.15	0.04 **
Talk about personal problem	1.26	0.09 *
Constant	4.11	0.98

Note: Models are weighted and adjusted for AddHealth's complex survey design.

^***p<.001;

^**p<.01;

^*p<.05

Table A7.

Parameters From Multinomial Logistic Regression Models Predicting Visits to Mothers by Respondent Intermarriage and Gender (N= 4,712)

	OR	SE B
Omitted: Less than weekly
Weekly
Respondent's traits
Intermarriage	0.91	0.15
Female	1.42	0.10 **
Intermarriage × Female	0.98	0.23
Race/ethnicity (/NH White)
NH Black	0.95	0.17
NH Asian	1.11	0.37
Hispanic (all races)	1.38	0.21
Child of immigrants	1.23	0.17
Age	0.97	0.03
Parent	1.18	0.11
Education (/< High school)
High school/GED	0.68	0.23
Some college	0.66	0.22
College or more	0.45	0.23 **
Labor force status (/not employed)
Part-time	1.64	0.16 **
Full-time	1.44	0.13 **
Mother's and household's traits
Mother's education at W1 (/< 8 grade)
8<grade<12	0.77	0.27
High school/GED	0.93	0.24
Some college	0.83	0.28
College or more	0.63	0.26
Family type at W1 (/both bio parents married)
Bio mother and stepfather	0.59	0.16 **
Bio mother only	0.66	0.10 ***
Relationship to mother at W1
Emotional closeness	1.27	0.09 *
Frequency of interaction	0.96	0.05
Talk about personal problem	0.88	0.08
Constant	2.19	0.93
Daily
Respondent's traits
Intermarriage	0.83	0.21
Female	2.40	0.14 ***
Intermarriage × Female	1.09	0.33
Race/ethnicity (/NH White)
NH Black	1.55	0.14 **
NH Asian	1.27	0.41
Hispanic (all races)	1.48	0.23
Child of immigrants	0.77	0.23
Age	1.01	0.04
Parent	2.28	0.15 ***
Education (/< High school)
High school/GED	0.68	0.28
Some college	0.52	0.26 *
College or more	0.23	0.28 ***
Labor force status (/not employed)
Part-time	1.56	0.18 *
Full-time	1.51	0.18 *
Mother's and household's traits
Mother's education at W1 (/< 8 grade)
8<grade<12	0.61	0.31
High school/GED	0.56	0.30
Some college	0.46	0.33 *
College or more	0.29	0.32 ***
Family type at W1 (/both bio parents married)
Bio mother and stepfather	0.52	0.17 ***
Bio mother only	0.58	0.15 ***
Relationship to mother at W1
Emotional closeness	1.56	0.13 **
Frequency of interaction	1.04	0.06
Talk about personal problem	0.85	0.11
Constant	0.27	1.12

Note: Models are weighted and adjusted for AddHealth's complex survey design.

^***p<.001;

^**p<.01;

^*p<.05

Table A8.

Parameters From Multinomial Logistic Regression Models Predicting Phone/Email Contact with Mothers by Respondent Intermarriage and Gender (N= 4,712)

	OR	SE B
Omitted: Less than weekly
Weekly
Respondent's Traits
Intermarriage	0.94	0.22
Female	2.65	0.20***
Intermarriage × Female	0.44	0.38 *
Race/ethnicity (/NH White)
NH Black	1.36	0.24
NH Asian	1.83	0.38
Hispanic (all races)	2.46	0.32 **
Child of immigrants	0.83	0.27
Age	1.02	0.04
Parent	0.78	0.13
Education (/< High school)
High school/GED	1.17	0.29
Some college	1.67	0.31
College or more	1.88	0.31 *
Labor force status (/not employed)
Part-time	0.69	0.28
Full-time	1.47	0.21
Mother's and household's traits
Mother's education at W1 (/< 8 grade)
8<grade<12	1.11	0.46
High school/GED	0.88	0.43
Some college	0.83	0.45
College or more	1.20	0.43
Family type at W1 (/both bio parents married)
Bio mother and stepfather	0.80	0.23
Bio mother only	0.61	0.15 **
Relationship to mother at W1
Emotional closeness	1.44	0.15*
Frequency of interaction	1.11	0.07
Talk about personal problem	0.92	0.14
Constant	0.59	1.53
Daily
Respondent's Traits
Intermarriage	1.01	0.25
Female	11.20	0.20 ***
Intermarriage × Female	0.45	0.41
Race/ethnicity (/NH White)
NH Black	2.41	0.26 **
NH Asian	1.00	0.42
Hispanic (all races)	2.35	0.30 **
Child of immigrants	0.73	0.25
Age	1.00	0.05
Parent	1.09	0.14
Education (/< High school)
High school/GED	0.84	0.34
Some college	1.15	0.31
College or more	0.80	0.31
Labor force status (/not employed)
Part-time	0.55	0.27 *
Full-time	1.26	0.21
Mother's and household's traits
Mother's education at W1 (/< 8 grade)
8<grade<12	1.29	0.48
High school/GED	0.94	0.40
Some college	0.86	0.41
College or more	0.85	0.41
Family type at W1 (/both bio parents married)
Bio mother and stepfather	0.60	0.26 *
Bio mother only	0.52	0.18 ***
Relationship to mother at W1
Emotional closeness	2.71	0.18 ***
Frequency of interaction	1.19	0.07 *
Talk about personal problem	0.94	0.14
Constant	0.64	1.63

Note: Models are weighted and adjusted for AddHealth's complex survey design.

^***p<.001;

^**p<.01;

^*p<.05